Google Data Analytics Certificate Capstone (Bellabeat Case Study)
Gabriel Fernandez
# Set default options for code chunks
knitr::opts_chunk$set(
echo = TRUE, # Display R code and its output
comment=NA, # Suppress code comments in output
warning = FALSE, # Suppress warning messages
fig.align='center', # Align figures in the center
eval = TRUE # Evaluate R code
)
Prepare and preprocess phase
# Import libraries
library(tidyverse) # includes ggplot2
library(skimr) # provides a compact and informative summary of your dataframe or dataset
library(lubridate)
library(janitor) # set of utility functions for data cleaning and data frame tidying tasks
library(RColorBrewer) # Color palettes for data visualization
library(ggcorrplot) # Visualize correlation matrices using ggplot2
library(scales) # formatting and transforming data for visualizations
# display.brewer.all(colorblindFriendly = TRUE)Load datasets
These datasets originate from a survey distributed on Amazon Mechanical Turk from 03.12.2016 to 05.12.2016. They include personal tracker data from 30 Fitbit users, covering physical activity, heart rate, and sleep monitoring, with differentiation based on Fitbit types and user behavior.
- Metadata: Fitbit data dictionary
# Clean environment
rm(list = ls())
daily_activity <-
read_csv("original_data/dailyActivity_merged.csv",
trim_ws = TRUE,
show_col_types = FALSE
)
daily_sleep <- read_csv("original_data/sleepDay_merged.csv",
trim_ws = TRUE,
show_col_types = FALSE
)
hourly_calories <-
read_csv("original_data/hourlyCalories_merged.csv",
trim_ws = TRUE,
show_col_types = FALSE
)
hourly_intensities <-
read_csv("original_data/hourlyIntensities_merged.csv",
trim_ws = TRUE,
show_col_types = FALSE
)
hourly_steps <-
read_csv("original_data/hourlySteps_merged.csv",
trim_ws = TRUE,
show_col_types = FALSE
)
minute_sleep <-
read_csv("original_data/minuteSleep_merged.csv",
trim_ws = TRUE,
show_col_types = FALSE
)
seconds_heartrate <-
read_csv("original_data/heartrate_seconds_merged.csv",
trim_ws = TRUE,
show_col_types = FALSE
)
weight_logs <-
read_csv("original_data/weightLogInfo_merged.csv",
trim_ws = TRUE,
show_col_types = FALSE
)
# Remove trailing spaces (trim_ws = TRUE)Clean datasets
Clean the daily_activity dataset
# Check daily_activity dataset before cleaning
glimpse(daily_activity)Rows: 940
Columns: 15
$ Id <dbl> 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 15…
$ ActivityDate <chr> "4/12/2016", "4/13/2016", "4/14/2016", "4/15/2016", "4/16/2016", "4/17/201…
$ TotalSteps <dbl> 13162, 10735, 10460, 9762, 12669, 9705, 13019, 15506, 10544, 9819, 12764, …
$ TotalDistance <dbl> 8.50, 6.97, 6.74, 6.28, 8.16, 6.48, 8.59, 9.88, 6.68, 6.34, 8.13, 9.04, 6.…
$ TrackerDistance <dbl> 8.50, 6.97, 6.74, 6.28, 8.16, 6.48, 8.59, 9.88, 6.68, 6.34, 8.13, 9.04, 6.…
$ LoggedActivitiesDistance <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
$ VeryActiveDistance <dbl> 1.88, 1.57, 2.44, 2.14, 2.71, 3.19, 3.25, 3.53, 1.96, 1.34, 4.76, 2.81, 2.…
$ ModeratelyActiveDistance <dbl> 0.55, 0.69, 0.40, 1.26, 0.41, 0.78, 0.64, 1.32, 0.48, 0.35, 1.12, 0.87, 0.…
$ LightActiveDistance <dbl> 6.06, 4.71, 3.91, 2.83, 5.04, 2.51, 4.71, 5.03, 4.24, 4.65, 2.24, 5.36, 3.…
$ SedentaryActiveDistance <dbl> 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.…
$ VeryActiveMinutes <dbl> 25, 21, 30, 29, 36, 38, 42, 50, 28, 19, 66, 41, 39, 73, 31, 78, 48, 16, 52…
$ FairlyActiveMinutes <dbl> 13, 19, 11, 34, 10, 20, 16, 31, 12, 8, 27, 21, 5, 14, 23, 11, 28, 12, 34, …
$ LightlyActiveMinutes <dbl> 328, 217, 181, 209, 221, 164, 233, 264, 205, 211, 130, 262, 238, 216, 279,…
$ SedentaryMinutes <dbl> 728, 776, 1218, 726, 773, 539, 1149, 775, 818, 838, 1217, 732, 709, 814, 8…
$ Calories <dbl> 1985, 1797, 1776, 1745, 1863, 1728, 1921, 2035, 1786, 1775, 1827, 1949, 17…# Check missing values and duplicates
cat(
"\n",
"Missing values:",
sum(is.na(daily_activity)),
"\n",
"Duplicate values:",
sum(duplicated(daily_activity)),
"\n",
"Unique Ids:",
n_distinct(daily_activity$Id)
)
Missing values: 0
Duplicate values: 0
Unique Ids: 33Let us clean:
Change column names to lower case because R is case sensitive.
Change “Id” from double to a character because the number represents a category.
Change “ActivityDate” from char to date.
# Clean daily_activity dataset
daily_activity <-
# Clean column names
clean_names(daily_activity) %>%
# Correct column types
mutate(id = as.character(id)) %>% # from double to chr
mutate(activity_date = as.Date(activity_date,
format = "%m/%d/%Y")) %>% # from chr to date
# Remove duplicate rows
distinct()
# Check daily_activity dataset after cleaning
glimpse(daily_activity)Rows: 940
Columns: 15
$ id <chr> "1503960366", "1503960366", "1503960366", "1503960366", "1503960366", "1…
$ activity_date <date> 2016-04-12, 2016-04-13, 2016-04-14, 2016-04-15, 2016-04-16, 2016-04-17,…
$ total_steps <dbl> 13162, 10735, 10460, 9762, 12669, 9705, 13019, 15506, 10544, 9819, 12764…
$ total_distance <dbl> 8.50, 6.97, 6.74, 6.28, 8.16, 6.48, 8.59, 9.88, 6.68, 6.34, 8.13, 9.04, …
$ tracker_distance <dbl> 8.50, 6.97, 6.74, 6.28, 8.16, 6.48, 8.59, 9.88, 6.68, 6.34, 8.13, 9.04, …
$ logged_activities_distance <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, …
$ very_active_distance <dbl> 1.88, 1.57, 2.44, 2.14, 2.71, 3.19, 3.25, 3.53, 1.96, 1.34, 4.76, 2.81, …
$ moderately_active_distance <dbl> 0.55, 0.69, 0.40, 1.26, 0.41, 0.78, 0.64, 1.32, 0.48, 0.35, 1.12, 0.87, …
$ light_active_distance <dbl> 6.06, 4.71, 3.91, 2.83, 5.04, 2.51, 4.71, 5.03, 4.24, 4.65, 2.24, 5.36, …
$ sedentary_active_distance <dbl> 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, …
$ very_active_minutes <dbl> 25, 21, 30, 29, 36, 38, 42, 50, 28, 19, 66, 41, 39, 73, 31, 78, 48, 16, …
$ fairly_active_minutes <dbl> 13, 19, 11, 34, 10, 20, 16, 31, 12, 8, 27, 21, 5, 14, 23, 11, 28, 12, 34…
$ lightly_active_minutes <dbl> 328, 217, 181, 209, 221, 164, 233, 264, 205, 211, 130, 262, 238, 216, 27…
$ sedentary_minutes <dbl> 728, 776, 1218, 726, 773, 539, 1149, 775, 818, 838, 1217, 732, 709, 814,…
$ calories <dbl> 1985, 1797, 1776, 1745, 1863, 1728, 1921, 2035, 1786, 1775, 1827, 1949, …# Check missing values and duplicates after cleaning
cat("\n",
"Missing values:",
sum(is.na(daily_activity)),
"\n",
"Duplicate values:",
sum(duplicated(daily_activity)))
Missing values: 0
Duplicate values: 0# Let us print summary statistics to have a better idea of the dataset
daily_activity %>%
summary() id activity_date total_steps total_distance tracker_distance
Length:940 Min. :2016-04-12 Min. : 0 Min. : 0.000 Min. : 0.000
Class :character 1st Qu.:2016-04-19 1st Qu.: 3790 1st Qu.: 2.620 1st Qu.: 2.620
Mode :character Median :2016-04-26 Median : 7406 Median : 5.245 Median : 5.245
Mean :2016-04-26 Mean : 7638 Mean : 5.490 Mean : 5.475
3rd Qu.:2016-05-04 3rd Qu.:10727 3rd Qu.: 7.713 3rd Qu.: 7.710
Max. :2016-05-12 Max. :36019 Max. :28.030 Max. :28.030
logged_activities_distance very_active_distance moderately_active_distance light_active_distance
Min. :0.0000 Min. : 0.000 Min. :0.0000 Min. : 0.000
1st Qu.:0.0000 1st Qu.: 0.000 1st Qu.:0.0000 1st Qu.: 1.945
Median :0.0000 Median : 0.210 Median :0.2400 Median : 3.365
Mean :0.1082 Mean : 1.503 Mean :0.5675 Mean : 3.341
3rd Qu.:0.0000 3rd Qu.: 2.053 3rd Qu.:0.8000 3rd Qu.: 4.782
Max. :4.9421 Max. :21.920 Max. :6.4800 Max. :10.710
sedentary_active_distance very_active_minutes fairly_active_minutes lightly_active_minutes
Min. :0.000000 Min. : 0.00 Min. : 0.00 Min. : 0.0
1st Qu.:0.000000 1st Qu.: 0.00 1st Qu.: 0.00 1st Qu.:127.0
Median :0.000000 Median : 4.00 Median : 6.00 Median :199.0
Mean :0.001606 Mean : 21.16 Mean : 13.56 Mean :192.8
3rd Qu.:0.000000 3rd Qu.: 32.00 3rd Qu.: 19.00 3rd Qu.:264.0
Max. :0.110000 Max. :210.00 Max. :143.00 Max. :518.0
sedentary_minutes calories
Min. : 0.0 Min. : 0
1st Qu.: 729.8 1st Qu.:1828
Median :1057.5 Median :2134
Mean : 991.2 Mean :2304
3rd Qu.:1229.5 3rd Qu.:2793
Max. :1440.0 Max. :4900 - This summary helps us explore each attribute quickly. We notice that some attributes have a minimum value of zero (total_step, total_distance, calories). Let us explore this observation.
# Check where total_steps is zero
filter(daily_activity, total_steps == 0) - We found 77 observations where total_steps is zero. We should delete these observations so they do not affect our mean and median. If the total_step is zero, the person did not wear the Fitbit.
# Check where calories is zero
filter(daily_activity, calories == 0)# Check where total_distance is zero
filter(daily_activity, total_distance == 0)- From our inspection above, we can see that we just need to delete the entries where total_steps is zero and will take take care of the rest.
daily_activity_clean <-
filter(daily_activity,
total_steps != 0,
total_distance != 0,
calories != 0)
daily_activity_cleannames(daily_activity) [1] "id" "activity_date" "total_steps"
[4] "total_distance" "tracker_distance" "logged_activities_distance"
[7] "very_active_distance" "moderately_active_distance" "light_active_distance"
[10] "sedentary_active_distance" "very_active_minutes" "fairly_active_minutes"
[13] "lightly_active_minutes" "sedentary_minutes" "calories" # Check the attributes again
cat("Before deleting the entries\n\n")Before deleting the entriesselect(daily_activity,total_steps,total_distance,calories) %>%
summary() total_steps total_distance calories
Min. : 0 Min. : 0.000 Min. : 0
1st Qu.: 3790 1st Qu.: 2.620 1st Qu.:1828
Median : 7406 Median : 5.245 Median :2134
Mean : 7638 Mean : 5.490 Mean :2304
3rd Qu.:10727 3rd Qu.: 7.713 3rd Qu.:2793
Max. :36019 Max. :28.030 Max. :4900 cat("\n\n\n",
"\t\t vs",
"\n\n\n")
vs cat("After deleting the entries\n\n")After deleting the entriesselect(daily_activity_clean, total_steps, total_distance, calories) %>%
summary() total_steps total_distance calories
Min. : 8 Min. : 0.010 Min. : 52
1st Qu.: 4927 1st Qu.: 3.373 1st Qu.:1857
Median : 8054 Median : 5.590 Median :2220
Mean : 8329 Mean : 5.986 Mean :2362
3rd Qu.:11096 3rd Qu.: 7.905 3rd Qu.:2832
Max. :36019 Max. :28.030 Max. :4900 - We can see that the observation we removed affected our mean and median.
Clean the daily_sleep dataset
# Check daily_sleep dataset before cleaning
glimpse(daily_sleep)Rows: 413
Columns: 5
$ Id <dbl> 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 15039603…
$ SleepDay <chr> "4/12/2016 12:00:00 AM", "4/13/2016 12:00:00 AM", "4/15/2016 12:00:00 AM", "4/16…
$ TotalSleepRecords <dbl> 1, 2, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
$ TotalMinutesAsleep <dbl> 327, 384, 412, 340, 700, 304, 360, 325, 361, 430, 277, 245, 366, 341, 404, 369, …
$ TotalTimeInBed <dbl> 346, 407, 442, 367, 712, 320, 377, 364, 384, 449, 323, 274, 393, 354, 425, 396, …# Check missing values and duplicates
cat("\n",
"Missing values:",
sum(is.na(daily_sleep)),
"\n",
"Duplicate values:",
sum(duplicated(daily_sleep)),
"\n",
"Unique Ids:",
n_distinct(daily_sleep$Id)
)
Missing values: 0
Duplicate values: 3
Unique Ids: 24Let us clean:
Change column names to lower case because R is case sensitive
Change “Id” from double to a character because the number represents a category
Change “SleepDay” from char to date. Since the time component of this column is the same for each observation”12:00:00 AM”, we can remove it. This will helps us merged this dataset with daily_activity later
Delete duplicates (3 observations are duplicates)
# Clean daily_sleep dataset
daily_sleep_clean <-
# Clean column names
clean_names(daily_sleep) %>%
# Correct column types
mutate(id = as.character(id)) %>% # from double to chr
mutate(sleep_day = as.Date(sleep_day,
format = "%m/%d/%Y")) %>% # from chr to date
# Remove duplicate rows
distinct()
# Check clean daily_sleep dataset
glimpse(daily_sleep_clean)Rows: 410
Columns: 5
$ id <chr> "1503960366", "1503960366", "1503960366", "1503960366", "1503960366", "1503960…
$ sleep_day <date> 2016-04-12, 2016-04-13, 2016-04-15, 2016-04-16, 2016-04-17, 2016-04-19, 2016-…
$ total_sleep_records <dbl> 1, 2, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, …
$ total_minutes_asleep <dbl> 327, 384, 412, 340, 700, 304, 360, 325, 361, 430, 277, 245, 366, 341, 404, 369…
$ total_time_in_bed <dbl> 346, 407, 442, 367, 712, 320, 377, 364, 384, 449, 323, 274, 393, 354, 425, 396…# Check missing values and duplicates after cleaning
cat("\n",
"Missing values:",
sum(is.na(daily_sleep_clean)),
"\n",
"Duplicate values:",
sum(duplicated(daily_sleep_clean)))
Missing values: 0
Duplicate values: 0Clean the hourly datasets (hourly_calories, hourly_intensities, and hourly_steps)
# Check hourly_calories dataset before cleaning
glimpse(hourly_calories)Rows: 22,099
Columns: 3
$ Id <dbl> 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 15…
$ ActivityHour <chr> "4/12/2016 12:00:00 AM", "4/12/2016 1:00:00 AM", "4/12/2016 2:00:00 AM", "4/12/2016 3:…
$ Calories <dbl> 81, 61, 59, 47, 48, 48, 48, 47, 68, 141, 99, 76, 73, 66, 110, 151, 76, 83, 124, 104, 1…# Check missing values and duplicates
cat("\n",
"Missing values:",
sum(is.na(hourly_calories)),
"\n",
"Duplicate values:",
sum(duplicated(hourly_calories)))
Missing values: 0
Duplicate values: 0# Check hourly_intensities dataset before cleaning
glimpse(hourly_intensities)Rows: 22,099
Columns: 4
$ Id <dbl> 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 1503960366…
$ ActivityHour <chr> "4/12/2016 12:00:00 AM", "4/12/2016 1:00:00 AM", "4/12/2016 2:00:00 AM", "4/12/201…
$ TotalIntensity <dbl> 20, 8, 7, 0, 0, 0, 0, 0, 13, 30, 29, 12, 11, 6, 36, 58, 13, 16, 29, 39, 41, 31, 9,…
$ AverageIntensity <dbl> 0.333333, 0.133333, 0.116667, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0.…# Check missing values and duplicates
cat("\n",
"Missing values:",
sum(is.na(hourly_intensities)),
"\n",
"Duplicate values:",
sum(duplicated(hourly_intensities)))
Missing values: 0
Duplicate values: 0# Check hourly_steps dataset before cleaning
glimpse(hourly_steps)Rows: 22,099
Columns: 3
$ Id <dbl> 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 15…
$ ActivityHour <chr> "4/12/2016 12:00:00 AM", "4/12/2016 1:00:00 AM", "4/12/2016 2:00:00 AM", "4/12/2016 3:…
$ StepTotal <dbl> 373, 160, 151, 0, 0, 0, 0, 0, 250, 1864, 676, 360, 253, 221, 1166, 2063, 344, 489, 138…# Check missing values and duplicates
cat("\n",
"Missing values:",
sum(is.na(hourly_steps)),
"\n",
"Duplicate values:",
sum(duplicated(hourly_steps)))
Missing values: 0
Duplicate values: 0Join hourly datasets to create a hourly_actitvity dataset
- These datasets shared the same Id and Activity_hour, let us join them into a new dataset (hourly_activity) before we clean them.
# Join the hourly datasets (hourly_calories, hourly_intensities, and hourly_steps)
hourly_activity <-
inner_join(hourly_calories,
hourly_intensities,
by = c("Id", "ActivityHour"))
hourly_activity <-
inner_join(hourly_activity, hourly_steps, by = c("Id", "ActivityHour"))# Check hourly_activity dataset before cleaning
glimpse(hourly_activity)Rows: 22,099
Columns: 6
$ Id <dbl> 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 1503960366…
$ ActivityHour <chr> "4/12/2016 12:00:00 AM", "4/12/2016 1:00:00 AM", "4/12/2016 2:00:00 AM", "4/12/201…
$ Calories <dbl> 81, 61, 59, 47, 48, 48, 48, 47, 68, 141, 99, 76, 73, 66, 110, 151, 76, 83, 124, 10…
$ TotalIntensity <dbl> 20, 8, 7, 0, 0, 0, 0, 0, 13, 30, 29, 12, 11, 6, 36, 58, 13, 16, 29, 39, 41, 31, 9,…
$ AverageIntensity <dbl> 0.333333, 0.133333, 0.116667, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0.…
$ StepTotal <dbl> 373, 160, 151, 0, 0, 0, 0, 0, 250, 1864, 676, 360, 253, 221, 1166, 2063, 344, 489,…# Check missing values and duplicates
cat("\n",
"Missing values:",
sum(is.na(hourly_activity)),
"\n",
"Duplicate values:",
sum(duplicated(hourly_activity)))
Missing values: 0
Duplicate values: 0Let us clean:
Change column names to lower case because R is case sensitive.
Change “Id” from double to a character because the number represents a category.
Change “ActivityHour” from char to datetime.
Note:The default timezone is UTC.
# Clean hourly_activity dataset
hourly_activity_clean <-
# Clean column names
clean_names(hourly_activity) %>%
# Correct column types
mutate(id = as.character(id)) %>% # from double to chr
mutate(activity_hour = as_datetime(activity_hour,
format = "%m/%d/%Y %I:%M:%S %p")) %>% # chr to datetime
# Remove duplicate rows
distinct()
# Check clean daily_activity dataset
glimpse(hourly_activity_clean)Rows: 22,099
Columns: 6
$ id <chr> "1503960366", "1503960366", "1503960366", "1503960366", "1503960366", "1503960366…
$ activity_hour <dttm> 2016-04-12 00:00:00, 2016-04-12 01:00:00, 2016-04-12 02:00:00, 2016-04-12 03:00:…
$ calories <dbl> 81, 61, 59, 47, 48, 48, 48, 47, 68, 141, 99, 76, 73, 66, 110, 151, 76, 83, 124, 1…
$ total_intensity <dbl> 20, 8, 7, 0, 0, 0, 0, 0, 13, 30, 29, 12, 11, 6, 36, 58, 13, 16, 29, 39, 41, 31, 9…
$ average_intensity <dbl> 0.333333, 0.133333, 0.116667, 0.000000, 0.000000, 0.000000, 0.000000, 0.000000, 0…
$ step_total <dbl> 373, 160, 151, 0, 0, 0, 0, 0, 250, 1864, 676, 360, 253, 221, 1166, 2063, 344, 489…# Check missing values and duplicates after cleaning
cat("\n",
"Missing values:",
sum(is.na(hourly_activity_clean)),
"\n",
"Duplicate values:",
sum(duplicated(hourly_activity_clean)))
Missing values: 0
Duplicate values: 0# as_datetime() converts with default timezone = "UTC"Clean the minute_sleep dataset
# Check minute_sleep dataset before cleaning
glimpse(minute_sleep)Rows: 188,521
Columns: 4
$ Id <dbl> 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 1503960366, 150396036…
$ date <chr> "4/12/2016 2:47:30 AM", "4/12/2016 2:48:30 AM", "4/12/2016 2:49:30 AM", "4/12/2016 2:50:30 AM…
$ value <dbl> 3, 2, 1, 1, 1, 1, 1, 2, 2, 2, 3, 3, 3, 3, 3, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, …
$ logId <dbl> 11380564589, 11380564589, 11380564589, 11380564589, 11380564589, 11380564589, 11380564589, 11…# Check missing values and duplicates
cat("\n",
"Missing values:",
sum(is.na(minute_sleep)),
"\n",
"Duplicate values:",
sum(duplicated(minute_sleep)),
"\n",
"Unique Ids:",
n_distinct(minute_sleep$Id))
Missing values: 0
Duplicate values: 543
Unique Ids: 24Let us clean:
Change column names to lower case because R is case sensitive.
Change “Id” from double to a character because the number represents a category.
Change “date” from char to datetime.
Change “value” from double to factor. Value indicates the sleep state: 1 = asleep, 2 = restless, 3 = awake. For for details see: Fitbit data dictionary
Remove duplicate values: 543.
# Clean minute_sleep dataset
minute_sleep_clean <-
# Clean column names
clean_names(minute_sleep) %>%
# Correct column types
mutate(value = as.factor(value)) %>% # from double to chr
mutate(id = as.character(id)) %>% # from double to chr
mutate(date = as_datetime(date,
format = "%m/%d/%Y %I:%M:%S %p")) %>% # From chr to datetime
# Remove duplicate rows
distinct()
# Check clean daily_activity dataset
glimpse(minute_sleep_clean)Rows: 187,978
Columns: 4
$ id <chr> "1503960366", "1503960366", "1503960366", "1503960366", "1503960366", "1503960366", "1503960…
$ date <dttm> 2016-04-12 02:47:30, 2016-04-12 02:48:30, 2016-04-12 02:49:30, 2016-04-12 02:50:30, 2016-04…
$ value <fct> 3, 2, 1, 1, 1, 1, 1, 2, 2, 2, 3, 3, 3, 3, 3, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,…
$ log_id <dbl> 11380564589, 11380564589, 11380564589, 11380564589, 11380564589, 11380564589, 11380564589, 1…# Check missing values and duplicates after cleaning
cat("\n",
"Missing values:",
sum(is.na(minute_sleep_clean)),
"\n",
"Duplicate values:",
sum(duplicated(minute_sleep_clean)))
Missing values: 0
Duplicate values: 0Clean the seconds_heartrate dataset
# Check seconds_heartrate set before cleaning
glimpse(seconds_heartrate)Rows: 2,483,658
Columns: 3
$ Id <dbl> 2022484408, 2022484408, 2022484408, 2022484408, 2022484408, 2022484408, 2022484408, 202248440…
$ Time <chr> "4/12/2016 7:21:00 AM", "4/12/2016 7:21:05 AM", "4/12/2016 7:21:10 AM", "4/12/2016 7:21:20 AM…
$ Value <dbl> 97, 102, 105, 103, 101, 95, 91, 93, 94, 93, 92, 89, 83, 61, 60, 61, 61, 57, 54, 55, 58, 60, 5…# Check missing values and duplicates
cat(
"\n",
"Missing values:", sum(is.na(seconds_heartrate)),
"\n",
"Duplicate values:", sum(duplicated(seconds_heartrate))
)
Missing values: 0
Duplicate values: 0Let us clean:
Change column names to lower case because R is case sensitive.
Change “Id” from double to a character because the number represents a category.
Change “Time” from char to datetime and rename it date_time.
Rename “Value” to heart_rate.
For more details see: Fitbit data dictionary
# Clean seconds_heartrate dataset
seconds_heartrate_clean <-
# Clean column names
clean_names(seconds_heartrate) %>%
# Correct column types
mutate(id = as.character(id)) %>% # from double to chr
mutate(time = as_datetime(time,
format = "%m/%d/%Y %I:%M:%S %p")) %>% # from chr to datetime
# Rename columns
rename(date_time = time,
heart_rate = value) %>%
# Remove duplicate rows
distinct()
# Check clean daily_activity dataset
glimpse(seconds_heartrate_clean)Rows: 2,483,658
Columns: 3
$ id <chr> "2022484408", "2022484408", "2022484408", "2022484408", "2022484408", "2022484408", "202…
$ date_time <dttm> 2016-04-12 07:21:00, 2016-04-12 07:21:05, 2016-04-12 07:21:10, 2016-04-12 07:21:20, 201…
$ heart_rate <dbl> 97, 102, 105, 103, 101, 95, 91, 93, 94, 93, 92, 89, 83, 61, 60, 61, 61, 57, 54, 55, 58, …# Check missing values and duplicates after cleaning
cat("\n",
"Missing values:",
sum(is.na(seconds_heartrate_clean)),
"\n",
"Duplicate values:",
sum(duplicated(seconds_heartrate_clean)))
Missing values: 0
Duplicate values: 0# as_datetime() converts with default timezone = "UTC"Clean the weight_logs dataset
# Check weight_logs set before cleaning
glimpse(weight_logs)Rows: 67
Columns: 8
$ Id <dbl> 1503960366, 1503960366, 1927972279, 2873212765, 2873212765, 4319703577, 4319703577, …
$ Date <chr> "5/2/2016 11:59:59 PM", "5/3/2016 11:59:59 PM", "4/13/2016 1:08:52 AM", "4/21/2016 1…
$ WeightKg <dbl> 52.6, 52.6, 133.5, 56.7, 57.3, 72.4, 72.3, 69.7, 70.3, 69.9, 69.2, 69.1, 90.7, 62.5,…
$ WeightPounds <dbl> 115.9631, 115.9631, 294.3171, 125.0021, 126.3249, 159.6147, 159.3942, 153.6622, 154.…
$ Fat <dbl> 22, NA, NA, NA, NA, 25, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, …
$ BMI <dbl> 22.65, 22.65, 47.54, 21.45, 21.69, 27.45, 27.38, 27.25, 27.46, 27.32, 27.04, 27.00, …
$ IsManualReport <lgl> TRUE, TRUE, FALSE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, FALSE, TRUE…
$ LogId <dbl> 1.462234e+12, 1.462320e+12, 1.460510e+12, 1.461283e+12, 1.463098e+12, 1.460938e+12, …# Check missing values and duplicates
cat("\n",
"Missing values:",
sum(is.na(weight_logs)),
"\n",
"Duplicate values:",
sum(duplicated(weight_logs)))
Missing values: 65
Duplicate values: 0Let us clean:
Change column names to lower case because R is case sensitive.
Change “Id” from double to a character because the number represents a category.
Change “Date” from char to datetime and rename it date_time.
Change NA to 0 in the column “fat.”
# Clean weight_logs dataset
weight_logs_clean <-
# Clean column names
clean_names(weight_logs) %>%
# Correct column types
mutate(id = as.character(id)) %>% # from double to chr
mutate(date = as_datetime(date,
format = "%m/%d/%Y %I:%M:%S %p")) %>% # from chr to datetime
# Rename columns
rename(date_time = date) %>%
# Remove duplicate rows
distinct()
# Change NA to 0 in the column "fat"
weight_logs_clean$fat[is.na(weight_logs_clean$fat)] <- 0
# Check clean daily_activity dataset
glimpse(weight_logs_clean)Rows: 67
Columns: 8
$ id <chr> "1503960366", "1503960366", "1927972279", "2873212765", "2873212765", "4319703577"…
$ date_time <dttm> 2016-05-02 23:59:59, 2016-05-03 23:59:59, 2016-04-13 01:08:52, 2016-04-21 23:59:5…
$ weight_kg <dbl> 52.6, 52.6, 133.5, 56.7, 57.3, 72.4, 72.3, 69.7, 70.3, 69.9, 69.2, 69.1, 90.7, 62.…
$ weight_pounds <dbl> 115.9631, 115.9631, 294.3171, 125.0021, 126.3249, 159.6147, 159.3942, 153.6622, 15…
$ fat <dbl> 22, 0, 0, 0, 0, 25, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,…
$ bmi <dbl> 22.65, 22.65, 47.54, 21.45, 21.69, 27.45, 27.38, 27.25, 27.46, 27.32, 27.04, 27.00…
$ is_manual_report <lgl> TRUE, TRUE, FALSE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, TRUE, FALSE, TR…
$ log_id <dbl> 1.462234e+12, 1.462320e+12, 1.460510e+12, 1.461283e+12, 1.463098e+12, 1.460938e+12…# Check missing values and duplicates after cleaning
cat("\n",
"Missing values:",
sum(is.na(weight_logs_clean)),
"\n",
"Duplicate values:",
sum(duplicated(weight_logs_clean)))
Missing values: 0
Duplicate values: 0Distribution of ids across datasets
# Loop through each dataset and print th number o funiqu ids
datasets <- c(
"daily_activity_clean",
"daily_sleep_clean",
"hourly_activity_clean",
"minute_sleep_clean",
"seconds_heartrate_clean",
"weight_logs_clean"
)
results_df <- data.frame(Dataset = character(0), distinct_IDs = integer(0))
for (dataset_name in datasets) {
dataset <- get(dataset_name) # Retrieve the dataset by its name
distinct_ids <- length(unique(dataset$id)) # Calculate the number of distinct IDs
result_row <- data.frame(Dataset = dataset_name, distinct_IDs = distinct_ids)
results_df <- bind_rows(results_df, result_row)
}
sorted_results <- results_df %>% arrange(- distinct_IDs )
print(sorted_results)NA- Differences in the number of unique IDs between the datasets can imply discrepancies in data collection methods, data incompleteness, or differing levels of user engagement.
Export clean datasets
# To uncomment the following code, select all the lines and press shift + control + c on Mac
# write.csv(daily_activity_clean,
# "daily_activity_clean.csv",
# row.names = FALSE)
#
# write.csv(daily_sleep_clean,
# "daily_sleep_clean.csv",
# row.names = FALSE)
#
# write.csv(daily_sleep_clean,
# "hourly_activity_clean.csv",
# row.names = FALSE)
#
# write.csv(minute_sleep_clean,
# "minute_sleep_clean.csv",
# row.names = FALSE)
#
# write.csv(seconds_heartrate_clean,
# "seconds_heartrate_clean.csv",
# row.names = FALSE)
#
# write.csv(weight_logs_clean ,
# "weight_logs_clean .csv",
# row.names = FALSE)
Analyze phase: Exploratory data analysis
EDA for daily_activity_clean
str(daily_activity_clean)tibble [862 × 15] (S3: tbl_df/tbl/data.frame)
$ id : chr [1:862] "1503960366" "1503960366" "1503960366" "1503960366" ...
$ activity_date : Date[1:862], format: "2016-04-12" "2016-04-13" "2016-04-14" ...
$ total_steps : num [1:862] 13162 10735 10460 9762 12669 ...
$ total_distance : num [1:862] 8.5 6.97 6.74 6.28 8.16 ...
$ tracker_distance : num [1:862] 8.5 6.97 6.74 6.28 8.16 ...
$ logged_activities_distance: num [1:862] 0 0 0 0 0 0 0 0 0 0 ...
$ very_active_distance : num [1:862] 1.88 1.57 2.44 2.14 2.71 ...
$ moderately_active_distance: num [1:862] 0.55 0.69 0.4 1.26 0.41 ...
$ light_active_distance : num [1:862] 6.06 4.71 3.91 2.83 5.04 ...
$ sedentary_active_distance : num [1:862] 0 0 0 0 0 0 0 0 0 0 ...
$ very_active_minutes : num [1:862] 25 21 30 29 36 38 42 50 28 19 ...
$ fairly_active_minutes : num [1:862] 13 19 11 34 10 20 16 31 12 8 ...
$ lightly_active_minutes : num [1:862] 328 217 181 209 221 164 233 264 205 211 ...
$ sedentary_minutes : num [1:862] 728 776 1218 726 773 ...
$ calories : num [1:862] 1985 1797 1776 1745 1863 ...Univariate analysis for daily_activity_clean
Numerical variables
# Subset numeric columns
num_df <- select_if(daily_activity_clean, is.numeric)
# Identify numeric columns
colnames(num_df) [1] "total_steps" "total_distance" "tracker_distance"
[4] "logged_activities_distance" "very_active_distance" "moderately_active_distance"
[7] "light_active_distance" "sedentary_active_distance" "very_active_minutes"
[10] "fairly_active_minutes" "lightly_active_minutes" "sedentary_minutes"
[13] "calories"
# plotting all numerical variables
col_names <- colnames(num_df)
for (i in col_names) {
suppressWarnings(print(
ggplot(num_df, aes(num_df[[i]])) +
geom_histogram(
bins = 30,
color = "black",
fill = "gray",
aes(y = ..density..)
) +
geom_density(
color = "blue",
size = 1
) +
xlab(i) + ylab("Count") +
ggtitle(paste("Histogram and Density Plot of", i))
))
}
NA
NA
NA
NAObservations:
Many variables show a right-skewed distribution: a larger number of data values are located on the left side of the curve.
The variables total_steps, total_distance, tracker_distance have a similar distribution. We can explore their correlations later.
Since the distributions are not normal. The median is a better indicator of central tendency for the numerical variables in these dataset.
The variable “logged_activities_distance” and “sedentary_active_distance” might not provide useful information since most of the data points are zero. It seems that the users are not logging the distance frequently.
The following variables seem related. We will explore them further in the bivariate analysis section:
sedentary_minutes; sedentary_active_distance
lightly_active_minutes; light_active_distance
fairly_active_minutes; moderately_active_distance
very_active_minutes; very_active_distance
The variables calories and sedentary_minutes exhibit a multimodal distribution, indicating the presence of subpopulations within the data. In this dataset, gender could be a potential variable that would result in a bimodal distribution when examining histograms of calories and sedentary minutes. Unfortunately, the gender of the users is not provided, limiting our ability to confirm this hypothesis.
Categorical variables
# Subset numeric columns
select_if(daily_activity_clean, negate(is.numeric))NA# Check counts by id
ggplot(data=daily_activity_clean) +
geom_bar(mapping = aes (x= reorder(id, id,length)))+
xlab("id") +
coord_flip()
#reorder(id, id, length) takes the id variable, uses itself to determine the order,
# and uses the length() function to calculate the values used for ordering.
# Essentially, this reorders the levels of the id variable based on the length of their names.
#https://stackoverflow.com/a/9231857/15333580count_max_ratio <- daily_activity_clean %>%
count(id) %>%
rename(id = "id", count = "n") %>%
mutate(percent_of_max = count / max(count) * 100) %>%
arrange(desc(percent_of_max))
# Create bar graph with percentage of entries compared to maximum
ggplot(count_max_ratio, aes(x = reorder(id, percent_of_max), y = percent_of_max)) +
geom_bar(stat = "identity") +
xlab("ID") +
ylab("Percentage of Maximum Count") +
ggtitle("Count by ID and Percentage of Maximum Count") +
theme_bw() +
theme(plot.title = element_text(hjust = 0.5)) +
geom_hline(yintercept=50, color="orange", linewidth=1)+
geom_hline(yintercept=75, color="red", linewidth=1)+
coord_flip()
NA
NA# percent_of_max > 75%
percent_of_max_top_75 <- filter(count_max_ratio, percent_of_max >=75)
percent_of_max_top_75 # percent_of_max < 75
percent_of_max_under_75 <- filter(count_max_ratio, percent_of_max < 75)
percent_of_max_under_75 daily_activity_clean$activity_date %>% summary() Min. 1st Qu. Median Mean 3rd Qu. Max.
"2016-04-12" "2016-04-18" "2016-04-26" "2016-04-26" "2016-05-03" "2016-05-12" ggplot(data=daily_activity_clean , aes(x = activity_date)) +
geom_histogram(binwidth = 1, color = "black", fill = "lightblue") +
labs(x = "Activity Date", y = "Frequency", title = "Distribution of Activity Date")
- It appears that there is missing activity data towards the end of the available period, specifically in the beginning of May.
# Investigate if the missing activity data coincides with the absence of entries for certain user IDs.
ggplot(data=subset(daily_activity_clean, id %in% percent_of_max_top_75$id), aes(x = activity_date)) +
geom_histogram(binwidth = 1, color = "black", fill = "lightblue") +
labs(x = "Activity Date", y = "Frequency", title = "Distribution of Activity Date For IDs with Above 75% of Entries")
ggplot(data=subset(daily_activity_clean, id %in% percent_of_max_under_75$id), aes(x = activity_date)) +
geom_histogram(binwidth = 1, color = "black", fill = "lightblue") +
labs(x = "Activity Date", y = "Frequency", title = "Distribution of Activity Date For IDs with under 75% of Entries")
- Users with more than 75% of data consistently report activity dates, while those with less than 75% of data show a decline in reporting starting from the end of April. The decline in Activity Date seems to be primarily due to a lack of data reporting from some users during that period.
Bivariate analysis
Correlation between numerical variables
corr <- cor(select_if(daily_activity_clean, is.numeric))
ggcorrplot(corr,
hc.order = TRUE,
type = "lower",
lab = TRUE,
colors = c("firebrick", "white", "royalblue"),
lab_size = 4,
lab_col = "black",
title = "Correlation Between Numerical Variables")
#https://rdrr.io/github/microresearcher/MicroVis/man/ggcorrplot.htmlsedentary_minutes; sedentary_active_distance lightly_active_minutes;
light_active_distance
fairly_active_minutes; moderately_active_distance very_active_minutes;
very_active_distance
# Compute correlation matrix
corr_matrix <- corr
# Set the threshold for correlation
threshold <- 0.60
# Find pairs of highly correlated variables
high_cor_pairs <- which(abs(corr_matrix) > threshold & lower.tri(corr_matrix, diag = FALSE), arr.ind = TRUE)
# Extract the variable names and correlation coefficients for the correlated pairs
variable_names <- colnames(corr_matrix)
cor_values <- as.vector(corr_matrix[high_cor_pairs])
# Create a data frame to store the correlated pairs and their correlation coefficients
cor_data <- data.frame(
Variable1 = variable_names[high_cor_pairs[, 1]],
Variable2 = variable_names[high_cor_pairs[, 2]],
Correlation = cor_values
)
# Sort the correlated pairs by correlation coefficient in descending order
sorted_cor_data <- cor_data[order(-cor_data$Correlation), ]
# Remove the index
row.names(sorted_cor_data) <- NULL
# Display the sorted correlated variable pairs in the dataframe
print(sorted_cor_data)NATotal_distance, tracker_distance, and total steps are highly correlated, so we will retain only total distance and total steps as they provide similar information.
The following minute and distance types are correlated. Which indicates that they report different aspects of the same activity, this is time or distance:
- lightly_active_minutes and light_active_distance (corr = 0.85)
- fairly_active_minutes and moderately_active_distance (corr = 0.94)
- very_active_minutes and very_active_distance (corr = 0.82)
There is a moderately high correlation between the time spent during very active periods and the total number of steps/total distance:
- The correlation between very_active_minutes and total_distance is 0.68
- The correlation between very_active_minutes and total_steps is 0.66
There is a moderate correlation of 0.61 between the total duration of very active minutes and the estimated daily calories consumed.
There is a moderate correlation of 0.62 between the total distance covered and the estimated daily calories consumed.
There is a moderate correlation coefficient of 0.60 between the distance covered during light activity (light_active_distance) and the total number of steps taken (total_steps).
Scatterplots of selected highly correlated variables pairs (>0.60)
# List of correlated variable pairs
correlated_pairs <- list(c("total_steps", "total_distance"),
c("lightly_active_minutes", "light_active_distance"),
c("fairly_active_minutes", "moderately_active_distance"),
c("very_active_minutes", "very_active_distance"),
c("very_active_minutes", "total_distance"),
c("very_active_minutes", "total_steps"),
c("very_active_minutes", "calories"),
c("total_distance", "calories"),
c("light_active_distance", "total_steps"))
# Loop over each pair and create scatter plot
for (pair in correlated_pairs) {
var1 <- pair[1]
var2 <- pair[2]
# Calculate averages
avg_var1 <- mean(daily_activity_clean[[var1]], na.rm = TRUE)
avg_var2 <- mean(daily_activity_clean[[var2]], na.rm = TRUE)
# Create scatter plot using ggplot2 with aes()
print(ggplot(data = daily_activity_clean, aes(x = !!sym(var1), y = !!sym(var2))) +
geom_point() +
geom_vline(xintercept = avg_var1, linetype = "dashed", color = "red") +
geom_hline(yintercept = avg_var2, linetype = "dashed", color = "blue") +
xlab(var1) + ylab(var2) +
ggtitle(paste("Scatter Plot with Average Reference Lines of", var1, "vs", var2)))
}
User Behavior for the daily activity dataset
Total steps: Total number of steps taken.
# Create a boxplot for total_steps
boxplot(daily_activity_clean$total_steps,
main = "Boxplot of Total Steps",
ylab = "Total Steps")
# Calculate the median and standard deviation
median_value <- median(daily_activity_clean$total_steps)
std_dev <- round(sd(daily_activity_clean$total_steps),2)
# Identify outliers
outliers <- boxplot.stats(daily_activity_clean$total_steps)$out
# Count the number of outliers
num_outliers <- length(outliers)
# Create the legend label with median, standard deviation, and outlier count
legend_label <- paste("Median:", median_value,
"\nStandard Deviation:", std_dev,
"\nOutliers:", num_outliers)
# Add the legend with median, standard deviation, and outlier count
legend("topright", legend = legend_label, pch = "", col = "black", bty = "n", cex = 0.85)
# Steps averages by IDs
steps_df <- daily_activity_clean %>%
group_by(id) %>%
summarise(average_steps = mean(total_steps), median_steps =median(total_steps), n = n())
steps_df# Calculate percentages for the average column
at_least_10k_avg <- sum(steps_df$average_steps >= 10000) / nrow(steps_df) * 100
between_5K_10K_avg <- sum(steps_df$average_steps >= 5000 & steps_df$average_steps < 10000) / nrow(steps_df) * 100
below_5k_avg <- sum(steps_df$average_steps < 5000) / nrow(steps_df) * 100
# Calculate percentages for the median column
at_least_10k_med <- sum(steps_df$median_steps >= 10000) / nrow(steps_df) * 100
between_5K_10K_med <- sum(steps_df$median_steps >= 5000 & steps_df$median_steps < 10000) / nrow(steps_df) * 100
below_5k_med <- sum(steps_df$median_steps < 5000) / nrow(steps_df) * 100
# Create a data frame for the steps categories
percentage_steps_df<- data.frame(
Category = c("Below 5,000", "Between 5,000 and 10,000", "At least 10,000"),
Percentage_Average = round(c(below_5k_avg, between_5K_10K_avg, at_least_10k_avg)),
Percentage_Median = round(c(below_5k_med, between_5K_10K_med, at_least_10k_med)))
percentage_steps_dfNA# Convert Category to a factor with custom factor levels
percentage_steps_df$Category <- factor(percentage_steps_df$Category, levels = c("Below 5,000", "Between 5,000 and 10,000", "At least 10,000"))
# Create a bar plot using ggplot
ggplot(percentage_steps_df, aes(x = Category, y = Percentage_Average)) +
geom_bar(stat = "identity", fill = "blue") +
labs(x = "Average Total Steps", y = "Percentage of Users", title = "58% of Users Average 5,000-10,000 Step Daily",subtitle = "Only 21% Achieve the 10,000-Step Goal") +
geom_text(aes(label = paste0(Percentage_Average, "%")), vjust = -0.5, color = "black") +
ylim(0, 100) + theme_minimal() + theme(panel.grid = element_blank())
NA
NATotal Distance: Total kilometers tracked.
# Create a boxplot for total_distance
boxplot(daily_activity_clean$total_distance,
main = "Boxplot of Total Distance",
ylab = "Total Distance")
# Calculate the median and standard deviation
median_value <- median(daily_activity_clean$total_distance)
std_dev <- sd(daily_activity_clean$total_distance)
# Identify outliers
outliers <- boxplot.stats(daily_activity_clean$total_distance)$out
# Count the number of outliers
num_outliers <- length(outliers)
# Create the legend label with median, standard deviation, and outlier count
legend_label <- paste("Median:", round(median_value, 2),
"\nStandard Deviation:", round(std_dev, 2),
"\nOutliers:", num_outliers)
# Add the legend with median, standard deviation, and outlier count
legend("topright", legend = legend_label, pch = "", col = "black", bty = "n")
# Total distance by IDs
t_distance_df <- daily_activity_clean %>%
group_by(id) %>%
summarise(average_t_distance = mean(total_distance ), median_t_distance =median(total_distance), n = n())
t_distance_df# Calculate percentages for the average column
at_least_10_avg<- sum(t_distance_df$average_t_distance>= 10) / nrow(t_distance_df) * 100
between_5_10_avg <- sum(t_distance_df$average_t_distance >= 5 & t_distance_df$average_t_distance < 10) / nrow(t_distance_df) * 100
below_5_avg <- sum(t_distance_df$average_t_distance < 5) / nrow(t_distance_df) * 100
# Create a data frame for the distance categories
percentage_t_distance_df<- data.frame(
Category = c("Below 5 km", "Between 5 and 10 km", "At least 10 km"),
Percentage_Average = round(c(below_5_avg, between_5_10_avg , at_least_10_avg)))
percentage_t_distance_df
# Convert Category to a factor with custom factor levels
percentage_t_distance_df$Category <- factor(percentage_t_distance_df$Category, levels = c("Below 5 km", "Between 5 and 10 km", "At least 10 km"))
# Create a bar plot using ggplot
ggplot(percentage_t_distance_df, aes(x = Category, y = Percentage_Average)) +
geom_bar(stat = "identity", fill = "pink") +
labs(x = "Average Total Distance", y = "Percentage of Users", title = "55% of Users Average 5-10 Kilometers Daily",subtitle = "10,000 steps is approximately equal to covering 5 miles (or 8 kilometers)") +
geom_text(aes(label = paste0(Percentage_Average, "%")), vjust = -0.5, color = "black") +
ylim(0, 100) + theme_minimal() +theme(panel.grid = element_blank())
Sedentary Minutes: Total minutes spent in sedentary activity.
# Create a boxplot for sedentary_minutes
boxplot(daily_activity_clean$sedentary_minutes,
main = "Boxplot of Sedentary Minutes",
ylab = "Sedentary Minutes")
# Calculate the median and standard deviation
median_value <- median(daily_activity_clean$sedentary_minutes)
std_dev <- sd(daily_activity_clean$sedentary_minutes)
# Identify outliers
outliers <- boxplot.stats(daily_activity_clean$sedentary_minutes)$out
# Count the number of outliers
num_outliers <- length(outliers)
# Create the legend label with median, standard deviation, and outlier count
legend_label <- paste("Median:", round(median_value, 2),
"\nStandard Deviation:", round(std_dev, 2),
"\nOutliers:", num_outliers)
# Add the legend with median, standard deviation, and outlier count
legend("topright", legend = legend_label, pch = "", col = "black", bty = "n", cex = 0.80)
NA
NA- These are high values for sedentary minutes. For instance, 1020 minutes is equivalent to 17 hours, and 1400 minutes is equivalent to 24 hours. After performing a quick search, it seems that the Fitbit uses 1400 as default for sedentary minutes when the device is not worn and it includes the sleeping time. SedentaryMinutes is total minutes spent in sedentary activity according to the data dictionary. See meta data section. Therefore, we need to substract the times sleeping to obtain an more accurate estimate of daily sedentary minutes.
# Check sedentary_minutes stats
daily_activity_clean$sedentary_minutes %>% summary() Min. 1st Qu. Median Mean 3rd Qu. Max.
0.0 721.2 1020.5 955.2 1189.0 1440.0 outliers[1] 2 13 0# Count entries where sedentary minutes equal 1440
count_1440 <- sum(daily_activity_clean$sedentary_minutes == 1440)
# Output the count
count_1440[1] 7# Remove rows with sedentary minutes equal to the default value (1440) and outliers
daily_activity_clean <- filter(daily_activity_clean, !(sedentary_minutes %in% c(0, 2, 13, 1440)))
# Rename the column
daily_sleep_clean <- rename(daily_sleep_clean, activity_date = sleep_day)
# Join the datasets
joined_activity_sleep <- inner_join(daily_activity_clean, daily_sleep_clean, by = c("id", "activity_date"))
# Check missing values and duplicates
cat(
"\n",
"Missing values:",
sum(is.na(joined_activity_sleep )),
"\n",
"Duplicate values:",
sum(duplicated(joined_activity_sleep )),
"\n",
"Unique Ids:",
n_distinct(joined_activity_sleep $id)
)
Missing values: 0
Duplicate values: 0
Unique Ids: 24# Create a derived column for sedentary minutes that does not include sleep time
joined_activity_sleep <- joined_activity_sleep %>%
mutate(
sedentary_min_awake = sedentary_minutes - total_minutes_asleep,
sedentary_hours_awake = sedentary_min_awake / 60,
sedentary_percentage_diff = (sedentary_minutes - sedentary_min_awake) / sedentary_minutes * 100
)
# Let us check the percentage difference of sedentary_minutes and the new column "sedentary_min_awake
# Create a boxplot for sedentary_percentage_diff
boxplot(joined_activity_sleep$sedentary_percentage_diff,
main = "Boxplot of Sedentary Percentage Difference",
ylab = "Sedentary Percentage Difference")
# Calculate the median and standard deviation
median_value <- median(joined_activity_sleep$sedentary_percentage_diff)
std_dev <- sd(joined_activity_sleep$sedentary_percentage_diff)
# Identify outliers
outliers <- boxplot.stats(joined_activity_sleep$sedentary_percentage_diff)$out
# Count the number of outliers
num_outliers <- length(outliers)
# Create the legend label with median, standard deviation, and outlier count
legend_label <- paste("Median:", round(median_value, 2),
"\nStandard Deviation:", round(std_dev, 2),
"\nOutliers:", num_outliers)
# Add the legend with median, standard deviation, and outlier count
legend("topright", legend = legend_label, pch = "", col = "black", bty = "n", cex = 0.80)
- The sedentary percentage difference has a median value of 59.95%, indicating a significant distinction between sedentary_minutes and sedentary_min_awake. This suggest that the original column “sedentary_minutes” included the time asleep.
# Create a boxplot for sedentary_min_awake
boxplot(joined_activity_sleep$sedentary_min_awake,
main = "Boxplot of Sedentary Minutes Awake",
ylab = "Sedentary Minutes Awake")
# Calculate the median and standard deviation
median_value <- median(joined_activity_sleep$sedentary_min_awake)
std_dev <- sd(joined_activity_sleep$sedentary_min_awake)
# Identify outliers
outliers <- boxplot.stats(joined_activity_sleep$sedentary_min_awake)$out
# Count the number of outliers
num_outliers <- length(outliers)
# Create the legend label with median, standard deviation, and outlier count
legend_label <- paste("Median:", round(median_value, 2),
"\nStandard Deviation:", round(std_dev, 2),
"\nOutliers:", num_outliers)
# Add the legend with median, standard deviation, and outlier count
legend("topright", legend = legend_label, pch = "", col = "black", bty = "n", cex = 0.80)
- Observation: There appears to be an inconsistency in the data. The sedentary_minutes value is smaller than the total_minutes_asleep value, which is unexpected.
# Count the number of cases where sedentary_minutes is smaller than total_minutes_asleep
count <- sum(joined_activity_sleep$sedentary_minutes < joined_activity_sleep$total_minutes_asleep)
# Print the count
count[1] 42# Subset the dataset
subset_data <- joined_activity_sleep[joined_activity_sleep$sedentary_minutes < joined_activity_sleep$total_minutes_asleep, ]
# View the subset data
subset_dataNA# Check column names of the subset data
subset_data %>%
select(sedentary_minutes, total_minutes_asleep, sedentary_min_awake, calories,id, activity_date, total_steps, total_distance, very_active_minutes )dim(subset_data)[1] 42 21dim(joined_activity_sleep)[1] 408 21# Use anti_join() to return a new dataset that includes all rows from the first dataset except for the rows that have a match in the second dataset.
clean_subset<- anti_join(joined_activity_sleep, subset_data)Joining with `by = join_by(id, activity_date, total_steps, total_distance, tracker_distance, logged_activities_distance, very_active_distance, moderately_active_distance, light_active_distance, sedentary_active_distance, very_active_minutes, fairly_active_minutes, lightly_active_minutes, sedentary_minutes, calories, total_sleep_records, total_minutes_asleep, total_time_in_bed, sedentary_min_awake, sedentary_hours_awake, sedentary_percentage_diff)`dim(clean_subset)[1] 366 21# Create a boxplot for sedentary_min_awake
boxplot(clean_subset$sedentary_min_awake,
main = "Boxplot of Sedentary Minutes Awake",
ylab = "Sedentary Minutes Awake")
# Calculate the median and standard deviation
median_value <- median(clean_subset$sedentary_min_awake)
std_dev <- sd(clean_subset$sedentary_min_awake)
# Identify outliers
outliers <- boxplot.stats(clean_subset$sedentary_min_awake)$out
# Count the number of outliers
num_outliers <- length(outliers)
# Create the legend label with median, standard deviation, and outlier count
legend_label <- paste("Median:", round(median_value, 2),
"\nStandard Deviation:", round(std_dev, 2),
"\nOutliers:", num_outliers)
# Add the legend with median, standard deviation, and outlier count
legend("topright", legend = legend_label, pch = "", col = "black", bty = "n", cex = 0.80)
Observation:
- By eliminating negative values from “sedentary_min_awake,” the resulting values now reflect a more realistic scenario.
# Total sedentary minutes awake by IDs
t_sedentary_df <- clean_subset %>%
group_by(id) %>%
summarise(average_sedentary_min_awake = mean(sedentary_min_awake),
median_sedentary_min_awake = median(sedentary_min_awake), n = n())
t_sedentary_dfNAdataset <- t_sedentary_df
column <- "average_sedentary_min_awake"
new_categories <- c("Below 200 minutes", "Between 200 and 400 minutes", "At least 400 minutes")
# Calculate percentages for the average column
below_200_avg <- sum(dataset[[column]] < 200) / nrow(dataset) * 100
between_200_400_avg <- sum(dataset[[column]] >= 200 & dataset[[column]] <= 400) / nrow(dataset) * 100
at_least_400_avg <- sum(dataset[[column]] >= 400) / nrow(dataset) * 100
# Create a data frame for the categories
percentage_sedentary_awake_df <- data.frame(
Category = new_categories,
Percentage_Average = round(c(below_200_avg, between_200_400_avg, at_least_400_avg))
)
# Convert Category to a factor with custom factor levels
percentage_sedentary_awake_df$Category <- factor(percentage_sedentary_awake_df$Category, levels = new_categories)
percentage_sedentary_awake_dfNA
NA
NA# Create a bar plot using ggplot
ggplot(percentage_sedentary_awake_df, aes(x = Category, y = Percentage_Average)) +
geom_bar(stat = "identity", fill = "gray") +
labs(x = "Average Total Sedentary Min Awake", y = "Percentage of Users",
title = "48% of Users Have an Average of at Least 400 Daily Sedentary Minutes While Awake",
subtitle = "200 Minutes are 3 hours and 20 minutes; 400 min are 6 hours and 40 min") +
geom_text(aes(label = paste0(Percentage_Average, "%")), vjust = -0.5, color = "black") +
ylim(0, 100) +
theme_minimal() +
theme(panel.grid = element_blank(), plot.title = element_text(size = 12), plot.subtitle = element_text(size = 10))
In a representative sample of U.S. adults, over two-thirds spent 6 + hours/day sitting, and more than half did not meet the recommended 150 min/week of physical activity. The study discovered that prolonged sitting for 6+ hours/day was associated with higher body fat percentages. While exceeding 150 min/week of physical activity was linked to lower body fat percentages, achieving recommended activity levels may not fully offset the increased body fat from prolonged sitting.
Jingwen Liao, Min Hu, Kellie Imm, Clifton J. Holmes, Jie Zhu, Chao Cao, Lin Yang. Association of daily sitting time and leisure-time physical activity with body fat among U.S. adults. Journal of Sport and Health Science, 2022. ISSN 2095-2546. https://doi.org/10.1016/j.jshs.2022.10.001. (https://www.sciencedirect.com/science/article/pii/S2095254622001016)
Calories:Total estimated energy expenditure (in kilocalories).
# Create a boxplot for calories
boxplot(daily_activity_clean$calories,
main = "Boxplot of Calories",
ylab = "Calories")
# Calculate the median and standard deviation
median_value <- median(daily_activity_clean$calories)
std_dev <- round(sd(daily_activity_clean$calories),2)
# Identify outliers
outliers <- boxplot.stats(daily_activity_clean$calories)$out
# Count the number of outliers
num_outliers <- length(outliers)
# Create the legend label with median, standard deviation, and outlier count
legend_label <- paste("Median:", median_value,
"\nStandard Deviation:", std_dev,
"\nOutliers:", num_outliers)
# Add the legend with median, standard deviation, and outlier count
legend("topright", legend = legend_label, pch = "", col = "black", bty = "n", cex = 0.85)
outliers[1] 4552 4392 4501 4546 4900 4547 4398# Calories averages by IDs
calories_df <- daily_activity_clean %>%
group_by(id) %>%
summarise(average_calories = mean(calories), median_calories = median(calories))
calories_dfNA# Calculate percentages for the average column
below_1600_avg <- sum(calories_df$average_calories < 1600) / nrow(calories_df) * 100
between_1600_2200_avg <- sum(calories_df$average_calories >= 1600 & calories_df$average_calories < 2200) / nrow(calories_df) * 100
between_2200_3000_avg <- sum(calories_df$average_calories >= 2200 & calories_df$average_calories < 3000) / nrow(calories_df) * 100
at_least_3000_avg <- sum(calories_df$average_calories >= 3000) / nrow(calories_df) * 100
# Calculate percentages for the median column
below_1600_med <- sum(calories_df$median_calories < 1600) / nrow(calories_df) * 100
between_1600_2200_med <- sum(calories_df$median_calories >= 1600 & calories_df$median_calories < 2200) / nrow(calories_df) * 100
between_2200_3000_med <- sum(calories_df$median_calories >= 2200 & calories_df$median_calories < 3000) / nrow(calories_df) * 100
at_least_3000_med <- sum(calories_df$median_calories >= 3000) / nrow(calories_df) * 100
# Create a data frame for the calories categories
percentage_calories_df <- data.frame(
Category = c("Below 1,600", "Between 1,600 and 2,200", "Between 2,200 and 3,000", "At least 3,000"),
Percentage_Average = round(c(below_1600_avg, between_1600_2200_avg, between_2200_3000_avg, at_least_3000_avg)),
Percentage_Median = round(c(below_1600_med, between_1600_2200_med, between_2200_3000_med, at_least_3000_med))
)
# Convert Category to a factor with custom factor levels
percentage_calories_df$Category <- factor(percentage_calories_df$Category, levels = c("Below 1,600", "Between 1,600 and 2,200", "Between 2,200 and 3,000", "At least 3,000"))
percentage_calories_dfNA# Create a bar plot using ggplot
ggplot(percentage_calories_df, aes(x = Category, y = Percentage_Average)) +
geom_bar(stat = "identity", fill = "red") +
labs(x = "Calorie Categories", y = "Percentage of Users",
title = "42% of Users Have an Average Daily Calorie Expenditure Between 1,600 and 2,200.",
subtitle = "Most females require 1,600 to 2,200 calories per day, as per the Dietary Guidelines for Americans, 2020-2025") +
geom_text(aes(label = paste0(Percentage_Average, "%")), vjust = -0.5, color = "black") +
ylim(0, 100) +
theme_minimal() +
theme(panel.grid = element_blank(),
plot.title = element_text(size = 12),
plot.subtitle = element_text(size = 10))
“Females ages 19 through 30 require about 1,800 to 2,400 calories a day. Males in this age group have higher calorie needs of about 2,400 to 3,000 a day. Calorie needs for adults ages 31 through 59 are generally lower; most females require about 1,600 to 2,200 calories a day and males require about 2,200 to 3,000 calories a day.”
U.S. Department of Agriculture and U.S. Department of Health and Human Services. Dietary Guidelines for Americans, 2020-2025. 9th Edition. December 2020. Available at DietaryGuidelines.gov/
Intensity Minutes: Time spent in one of four intensity categories.
VeryActiveMinutes: Total minutes spent in very active activity
FairlyActiveMinutes: Total minutes spent in moderate activity
LightlyActiveMinutes: Total minutes spent in light activity
SedentaryMinutes: Total minutes spent in sedentary activity
activity_minutes_df <- daily_activity_clean %>%
group_by(id) %>%
summarise(
average_very_active_minutes = mean(very_active_minutes),
average_fairly_active_minutes = mean(fairly_active_minutes),
average_lightly_active_minutes = mean(lightly_active_minutes),
average_sedentary_minutes = mean(sedentary_minutes)
)
activity_minutes_dfNA
NA
# Define the custom order of legend items
custom_order <- c( "Very Active", "Fairly Active", "Lightly Active", "Sedentary")
# Create the stacked bar plot
ggplot(activity_minutes_df, aes(y = id)) +
geom_bar(aes(x = average_sedentary_minutes, fill = "Sedentary"), stat = "identity", width = 0.5) +
geom_bar(aes(x = average_lightly_active_minutes, fill = "Lightly Active"), stat = "identity", width = 0.5) +
geom_bar(aes(x = average_fairly_active_minutes, fill = "Fairly Active"), stat = "identity", width = 0.5) +
geom_bar(aes(x = average_very_active_minutes, fill = "Very Active"), stat = "identity", width = 0.5) +
xlab("Minutes") +
ylab("ID") +
ggtitle("Average Activity Minutes by ID") +
scale_fill_manual(name = "", values = c("Very Active" = "red", "Fairly Active" = "orange", "Lightly Active" = "lightgreen", "Sedentary" = "lightblue"), breaks = custom_order) +
theme_minimal() +
theme(legend.position = "bottom", panel.grid = element_blank()) 
NA
NA
NA
NA
NA
NA
NA
NA# Calculate the average for each column
averages <- colMeans(activity_minutes_df[, c("average_very_active_minutes",
"average_fairly_active_minutes",
"average_lightly_active_minutes",
"average_sedentary_minutes")])
# Calculate the total average
total_average <- sum(averages)
# Calculate the proportions
proportions <- averages / total_average
# Create the new dataframe with modified row names
overall_average_df<- data.frame(Average = averages,
Percentage = proportions * 100)
# Modify the row names
row_names <- c("Very Active", "Fairly Active", "Lightly Active", "Sedentary")
row.names(overall_average_df) <- row_names
# Print the new dataframe
overall_average_dfNA
NA
ggplot(overall_average_df, aes(x = Percentage, y = reorder(row.names(overall_average_df), Percentage), fill = row.names(overall_average_df))) +
geom_bar(stat = "identity", width = 0.7, show.legend = FALSE) +
geom_text(aes(label = paste0(round(Percentage), "%")), hjust = -0.2, color = "black", size = 4) +
ylab("Minutes Intensity") +
xlab("Percentage") +
ggtitle("Users' Overall Average Intensity Minutes Consist Primarily of Sedentary and Lightly Active Time") +
scale_fill_manual(values = c("Very Active" = "red", "Fairly Active" = "orange", "Lightly Active" = "lightgreen", "Sedentary" = "lightblue")) +
scale_x_continuous(labels = NULL) +
theme_minimal() +
theme(legend.position = "none", panel.grid = element_blank(), axis.text.y = element_text(size = 10))
NA
NA“Analyzing each individual’s average calorie intake can provide insights into their individual dietary habits and patterns. By comparing the individual averages to the overall average, you can identify individuals who consume more or fewer calories compared to the group average. This comparison can help in understanding variations in calorie intake and potential factors influencing individual differences.”
# Define the custom order of legend items
custom_order <- c("Very Active", "Fairly Active", "Lightly Active", "Sedentary")
# Create the stacked horizontal bar chart
ggplot(overall_average_df, aes(x = Percentage, y = factor(1), fill = factor(row.names(overall_average_df), levels = custom_order))) +
geom_bar(stat = "identity", width = 0.7) +
xlab("Percentage") +
ylab("") +
ggtitle("Users' Overall Average Intensity Minutes Consist Primarily of Sedentary and Lightly Active Time") +
scale_fill_manual(
name = "",
values = c(
"Very Active" = "red",
"Fairly Active" = "orange",
"Lightly Active" = "lightgreen",
"Sedentary" = "lightblue"
),
breaks = custom_order
) +
guides(fill = guide_legend(reverse = TRUE)) + # Reverse the order of the legend
theme_minimal() +
theme(legend.position = "top",
panel.grid = element_blank(),
axis.text.y = element_blank(), # Remove the y-axis text
plot.title = element_text(size = 12, margin = margin(b = 20))) + # Adjust the title size and margin
geom_vline(xintercept = 97, color = "black", linetype = "dashed") +
annotate("text", x = 97, y = 1, label = " 97%", vjust = -5.5, hjust = 0.1)
NA
NAThese indicators provide insights into activity levels, sedentary behavior, and calorie burn. They can help track progress, set goals, and evaluate user behavior over time. Remember to consider the specific context and goals of your analysis to select and customize the most relevant KPIs for your use case. The context I will use is the guidelines for physical activity and diet for Americans:
- U.S. Department of Health and Human Services. (2019). Physical
Activity Guidelines for Americans (2nd ed.). Available at https://health.gov/sites/default/files/2019-09/Physical_Activity_Guidelines_2nd_edition.pdf
- U.S. Department of Agriculture and U.S. Department of Health and Human Services. Dietary Guidelines for Americans, 2020-2025. 9th Edition. December 2020. Available at DietaryGuidelines.gov/
EDA for daily_sleep_clean
str(daily_sleep_clean)tibble [410 × 5] (S3: tbl_df/tbl/data.frame)
$ id : chr [1:410] "1503960366" "1503960366" "1503960366" "1503960366" ...
$ activity_date : Date[1:410], format: "2016-04-12" "2016-04-13" "2016-04-15" ...
$ total_sleep_records : num [1:410] 1 2 1 2 1 1 1 1 1 1 ...
$ total_minutes_asleep: num [1:410] 327 384 412 340 700 304 360 325 361 430 ...
$ total_time_in_bed : num [1:410] 346 407 442 367 712 320 377 364 384 449 ...activity_date (sleep_day): Date on which the sleep event started.
total_sleep_records: Number of recorded sleep periods for that day. Includes naps > 60 min.
total_minutes_asleep: Total number of minutes classified as being “asleep”.
total_time_in_bed: Total minutes spent in bed, including asleep, restless, and awake, that occurred during a defined sleep record.
#Sanity check: Verify that the value of total_time_in_bed is greater than total_minutes_asleep, as we would expect.
daily_sleep_clean[daily_sleep_clean$total_time_in_bed < daily_sleep_clean$total_minutes_asleep,]Univariate analysis
numerical_cols <- daily_sleep_clean%>%
select_if(is.numeric)
# plotting all numerical variables
col_names <- colnames(numerical_cols )
for (i in col_names) {
suppressWarnings(print(
ggplot(numerical_cols , aes(numerical_cols [[i]])) +
geom_histogram(
bins = 30,
color = "black",
fill = "gray",
aes(y = ..density..)
) +
geom_density(
color = "blue",
size = 1
) +
xlab(i) + ylab("Count") +
ggtitle(paste("Histogram with Density Plot of", i))
))
}
Bivariate analysis
Correlation between numerical variables
# Correlation between numerical variables
corr <- cor(select_if(daily_sleep_clean, is.numeric))
ggcorrplot(corr,
hc.order = TRUE,
type = "lower",
lab = TRUE,
colors = c("firebrick", "white", "royalblue"),
lab_size = 4,
lab_col = "black",
title = "Correlation Between Numerical Variables")
NA
NAScatterplots of total_minutes_asleep vs total_time_in_bed
ggplot(data = daily_sleep_clean, aes(x = total_minutes_asleep, y = total_time_in_bed)) +
geom_point()
User Behavior for daily sleep dataset
frequency_table <- as.data.frame(table(daily_sleep_clean$total_sleep_records))
frequency_table$Percentage <- frequency_table$Freq / sum(frequency_table$Freq) * 100
ggplot(data = frequency_table, aes(x = Var1, y = Freq)) +
geom_bar(stat = "identity", fill = "steelblue") +
geom_text(aes(label = paste(Freq, " (", percent(Percentage / 100), ")", sep = "")),
hjust = 0.5, vjust = -0.4, color = "black") +
labs(x = "Total Sleep Records", y = "Frequency",
title = "Uncommon Napping: 89% of Sleep Records Indicate a Singular Sleep Period.",
subtitle = "Includes naps > 60 min.")+
theme_minimal() +
theme(panel.grid = element_blank(),
plot.title = element_text(size = 12),
plot.subtitle = element_text(size = 10, margin = margin(b = 20)))
NA
NA
NATotal minutes asleep
# Create a boxplot for total_minutes_asleep
boxplot(daily_sleep_clean$total_minutes_asleep,
main = "Boxplot of Total Minutes Asleep",
ylab = "Total Minutes Asleep")
# Calculate the median and standard deviation
median_value <- median(daily_sleep_clean$total_minutes_asleep)
std_dev <- round(sd(daily_sleep_clean$total_minutes_asleep), 2)
# Identify outliers
outliers <- boxplot.stats(daily_sleep_clean$total_minutes_asleep)$out
# Count the number of outliers
num_outliers <- length(outliers)
# Create the legend label with median, standard deviation, and outlier count
legend_label <- paste("Median:", median_value,
"\nStandard Deviation:", std_dev,
"\nOutliers:", num_outliers)
# Add the legend with median, standard deviation, and outlier count
legend("topright", legend = legend_label, pch = "", col = "black", bty = "n", cex = 0.85)
# Sleep duration averages by IDs with standard deviation and count (n)
sleep_df <- daily_sleep_clean %>%
group_by(id) %>%
summarise(average_sleep_minutes = mean(total_minutes_asleep),
standard_deviation_sleep_minutes = sd(total_minutes_asleep),
n = n())
sleep_dfNA
NA
NA# Drop ID "2320127002" due to insufficient data for computing mean and standard deviation.
sleep_df <- sleep_df %>%
filter(id != "2320127002")sleep_df # Calculate percentages for the average column
below_6_hours <- sum(sleep_df$average_sleep_minutes < 360) / nrow(sleep_df) * 100
between_6_7_hours <- sum(sleep_df$average_sleep_minutes >= 360 & sleep_df$average_sleep_minutes < 420) / nrow(sleep_df) * 100
at_least_7_hours <- sum(sleep_df$average_sleep_minutes >= 420) / nrow(sleep_df) * 100
# Create a data frame for the sleep duration categories
percentage_sleep_df <- data.frame(
Category = c("Below 6 hours", "Between 6 and 7 hours", "At least 7 hours"),
Percentage_Average = round(c(below_6_hours, between_6_7_hours, at_least_7_hours))
)
# Convert Category to a factor with custom factor levels
percentage_sleep_df$Category <- factor(percentage_sleep_df$Category, levels = c("Below 6 hours", "Between 6 and 7 hours", "At least 7 hours"))
percentage_sleep_dfstr(percentage_sleep_df)'data.frame': 3 obs. of 2 variables:
$ Category : Factor w/ 3 levels "Below 6 hours",..: 1 2 3
$ Percentage_Average: num 30 22 48
ggplot(percentage_sleep_df, aes(x = Category, y = Percentage_Average)) +
geom_bar(stat = "identity", fill = "purple") +
labs(x = "Average Sleep Duration", y = "Percentage of Users",
title = "52% of Users Get Less Than 7 Hours of Sleep on Average Daily") +
geom_text(aes(label = paste0(Percentage_Average, "%")), vjust = -0.5, color = "black") +
ylim(0, 100) +
theme_minimal() +
theme(panel.grid = element_blank(), plot.title = element_text(size = 12), plot.subtitle = element_text(size = 10))
Sleep duration consistency
#Error bars
# Convert average_sleep_minutes and standard_deviation_sleep_minutes to hours
sleep_df$average_sleep_hours <- sleep_df$average_sleep_minutes / 60
sleep_df$standard_deviation_sleep_hours <- sleep_df$standard_deviation_sleep_minutes / 60
# Create a bar plot for each 'id' with error bars representing standard deviation
ggplot(sleep_df, aes(x = id, y = average_sleep_hours)) +
geom_bar(stat = "identity", fill = "skyblue", color = "black") +
geom_errorbar(aes(ymin = average_sleep_hours - standard_deviation_sleep_minutes / 60,
ymax = average_sleep_hours + standard_deviation_sleep_minutes / 60),
width = 0.2, position = position_dodge(0.9), color = "black") +
labs(x = "ID", y = "Average Sleep Duration (hours)",
title = "Sleep Consistency: Average Sleep Duration with Error Bars",
subtitle = "Error bars represent the standard deviation around the mean.") +
theme_minimal() +
theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
geom_hline(yintercept = 7, linetype = "dashed", color = "red") +
scale_y_continuous(breaks = seq(0, 12, 1)) # Adjust the range as needed
# Calculate sleep duration averages and standard deviations in hours
sleep_df <- daily_sleep_clean %>%
group_by(id) %>%
summarise(n = n(),
average_sleep_hours = mean(total_minutes_asleep) / 60, # Convert minutes to hours
average_time_in_bed_hours = mean(total_time_in_bed) / 60,
standard_deviation_sleep_hours = sd(total_minutes_asleep) / 60,
standard_deviation_time_in_bed_hours = sd(total_time_in_bed) / 60,
) %>%
mutate(time_difference_hours = average_time_in_bed_hours - average_sleep_hours, # Calculate the time difference in hours
average_awake_in_bed_hours = time_difference_hours, # Rename column "awake_in_bed"
sd_awake_in_bed_hours = sd(time_difference_hours)) # Calculate SD for "awake_in_bed" in hours
sleep_dfNA
NA
NA# Drop ID "2320127002" due to insufficient data for computing mean and standard deviation.
sleep_df <- sleep_df %>%
filter(id != "2320127002")
dim(sleep_df)[1] 23 9create_boxplots_in_one_output <- function(data_frame, columns_to_analyze, decimal_places = 2) {
num_columns <- length(columns_to_analyze)
num_rows <- ceiling(num_columns / 2)
par(mfrow = c(num_rows, 2)) # Set the plotting layout
for (i in 1:num_columns) {
column_name <- columns_to_analyze[i]
boxplot(data_frame[[column_name]],
ylab = column_name)
median_value <- median(data_frame[[column_name]])
std_dev <- round(sd(data_frame[[column_name]]), decimal_places)
outliers <- boxplot.stats(data_frame[[column_name]])$out
num_outliers <- length(outliers)
legend_label <- paste("Median:", round(median_value, decimal_places),
"\nSD:", std_dev,
"\nOutliers:", num_outliers)
legend("topright", legend = legend_label, pch = "", col = "black", bty = "n", cex = 0.75)
}
par(mfrow = c(1, 1)) # Reset the plotting layout to default
}
# Columns to analyze
columns_to_analyze <- c("average_sleep_hours", "average_awake_in_bed_hours")
# Call the function to create boxplots in one output
create_boxplots_in_one_output(sleep_df, columns_to_analyze, decimal_places = 2)
# Columns to analyze
columns_to_analyze <- c("standard_deviation_sleep_hours", "sd_awake_in_bed_hours")
# Call the function
create_boxplots_in_one_output(sleep_df, columns_to_analyze, decimal_places = 2)
#Columns with outliers to remove
columns_with_outliers <- c("average_sleep_hours", "average_awake_in_bed_hours", "standard_deviation_sleep_hours")
# Function to remove outliers from a column
remove_outliers <- function(data, column_name) {
outlier_bounds <- boxplot.stats(data[[column_name]])$out
data_no_outliers<- data[!(data[[column_name]] %in% outlier_bounds), ]
return(data_no_outliers)
}
# Loop through each column and remove outliers
for (col in columns_with_outliers) {
sleep_df <- remove_outliers(sleep_df, col)
}
sleep_df# Check if outliers were removed
columns_to_analyze <- c("average_sleep_hours", "average_awake_in_bed_hours", "standard_deviation_sleep_hours")
# Call the function to create boxplots in one output
create_boxplots_in_one_output(sleep_df, columns_to_analyze, decimal_places = 2)
#Let us divide the users into irregular sleepers and regular sleepers. We will use the 75th percentile as the threshold to determine irregular sleepers. The rest will be considered regular sleepers.
# Define the Threshold (e.g., using the 75th percentile)
threshold <- quantile(sleep_df$standard_deviation_sleep_hours, 0.75)
# Create a new column "sleeper_type" based on the threshold
sleep_df$sleeper_type <- ifelse(sleep_df$standard_deviation_sleep_hours > threshold, "irregular", "regular")
sleep_df# sleep_type counts
table(sleep_df$sleeper_type)
irregular regular
4 13 sleep_df
color_options <- c("#E69F00", "#0072B2") # Blue: "#0072B2", Orange: "#E69F00"
# Function to create the violin plot for a given y-axis column
create_violin_plot <- function(data, x_axis_col, y_axis_col) {
ggplot(data, aes_string(x = x_axis_col, y = y_axis_col, fill = x_axis_col)) +
geom_violin(scale = "width", draw_quantiles = c(0.25, 0.5, 0.75), trim = FALSE) +
geom_boxplot(width = 0.1, fill = "white", color = "black") +
labs(x = "Sleeper Type", y = y_axis_col, title = paste("Comparison",x_axis_col,"for", y_axis_col)) +
scale_fill_manual(values = color_options) +
theme_minimal()
}
# Call the function to create the violin plots for each column
for (col in c("average_sleep_hours", "standard_deviation_sleep_hours", "average_awake_in_bed_hours","sd_awake_in_bed_hours")) {
plot <- create_violin_plot(sleep_df, "sleeper_type", col)
print(plot)
}
Observations:
Regular sleepers tend to have higher median average sleep hours compared to irregular sleepers.This suggests that individuals classified as regular sleepers are likely getting more sleep on average than those categorized as irregular sleepers.
Additionally, the spread of the “average_sleep_hours” for irregular sleepers appears to be wider, indicating more variability in their sleep duration. In contrast, the violin plot for regular sleepers shows a narrower spread, suggesting that their sleep duration is more consistent.
Regular sleepers exhibit a slightly higher median average awake-in-bed duration compared to irregular sleepers.
Summary: Regular sleepers get more sleep on average, have a more consistent sleep duration, and slightly higher median awake-in-bed duration than irregular sleepers.
EDA minute_sleep_clean
str(minute_sleep_clean)tibble [187,978 × 4] (S3: tbl_df/tbl/data.frame)
$ id : chr [1:187978] "1503960366" "1503960366" "1503960366" "1503960366" ...
$ date : POSIXct[1:187978], format: "2016-04-12 02:47:30" "2016-04-12 02:48:30" "2016-04-12 02:49:30" ...
$ value : Factor w/ 3 levels "1","2","3": 3 2 1 1 1 1 1 2 2 2 ...
$ log_id: num [1:187978] 1.14e+10 1.14e+10 1.14e+10 1.14e+10 1.14e+10 ...This data seems to come from the Classic Sleep Log (1 minute)
Value indicating the sleep state. 1 = asleep, 2 = restless, 3 = awake
For more detail check : Fitbit data dictionary
# Add labels to the velue column
minute_sleep_clean$value <- factor(minute_sleep_clean$value, levels = c("1", "2", "3"), labels = c("asleep", "restless", "awake"))minute_sleep_clean %>% summary() id date value log_id
Length:187978 Min. :2016-04-11 20:48:00.00 asleep :171960 Min. :1.137e+10
Class :character 1st Qu.:2016-04-19 02:48:00.00 restless: 14002 1st Qu.:1.144e+10
Mode :character Median :2016-04-26 21:48:00.00 awake : 2016 Median :1.150e+10
Mean :2016-04-26 13:31:23.11 Mean :1.150e+10
3rd Qu.:2016-05-03 23:47:00.00 3rd Qu.:1.155e+10
Max. :2016-05-12 09:56:00.00 Max. :1.162e+10
# Assuming "value" column represents total sleep records
frequency_table <- as.data.frame(table(minute_sleep_clean$value))
frequency_table$Percentage <- frequency_table$Freq / sum(frequency_table$Freq) * 100
ggplot(data = frequency_table, aes(x = Var1, y = Freq)) +
geom_bar(stat = "identity", fill = "#008080") +
geom_text(aes(label = paste(Freq, " (", percent(Percentage / 100), ")", sep = "")),
hjust = 0.5, vjust = -0.4, color = "black") +
labs(x = "Total Minutes Records", y = "Frequency",
title = "User Sleep States: 91% of Minutes Spent Asleep with Minimal Interruptions:",
subtitle = "Restlessness: 7.4% | Awake: 1.1%") +
theme_minimal() +
theme(panel.grid = element_blank(),
plot.title = element_text(size = 12),
plot.subtitle = element_text(size = 10, margin = margin(b = 20)))
EDA for hourly_activity_clean
str(hourly_activity_clean)tibble [22,099 × 6] (S3: tbl_df/tbl/data.frame)
$ id : chr [1:22099] "1503960366" "1503960366" "1503960366" "1503960366" ...
$ activity_hour : POSIXct[1:22099], format: "2016-04-12 00:00:00" "2016-04-12 01:00:00" "2016-04-12 02:00:00" ...
$ calories : num [1:22099] 81 61 59 47 48 48 48 47 68 141 ...
$ total_intensity : num [1:22099] 20 8 7 0 0 0 0 0 13 30 ...
$ average_intensity: num [1:22099] 0.333 0.133 0.117 0 0 ...
$ step_total : num [1:22099] 373 160 151 0 0 ...Calories integer Total number of estimated calories burned.
TotalIntensity: integer Value calculated by adding all the minute-level intensity values that occurred within the hour.
AverageIntensity: intensity state exhibited during that hour (TotalIntensity for that ActivityHour divided by 60)
StepTotal: Total number of steps taken.
For more detail check : Fitbit data dictionary
hourly_df <-hourly_activity_clean
# Extract "am" or "pm" from the activity_hour column
hourly_df$am_pm <- ifelse(format(hourly_df$activity_hour, "%p") == "AM", "am", "pm")
#Add a column for the hour
hourly_df$hour <- format(hourly_df$activity_hour, "%H")
# Group by id, hour, and am_pm and summarize the columns
summary_data <- hourly_df %>%
group_by(id, hour, am_pm) %>%
summarize(
calories_avg = mean(calories),
calories_max = max(calories),
calories_min = min(calories),
total_intensity_avg = mean(total_intensity),
total_intensity_max = max(total_intensity),
total_intensity_min = min(total_intensity),
average_intensity_avg = mean(average_intensity),
step_total_avg = mean(step_total),
observations_count = n()
)`summarise()` has grouped output by 'id', 'hour'. You can override using the `.groups` argument.
# Define colors for AM and PM
color_palette <- c("#FFA500", "#ADD8E6") # Orange for AM, Light Blue for PM
# Custom function to generate the bar plot with dynamically set y-axis limits
generate_bar_plot <- function(data, y_var, y_label, limit_factor) {
y_limit <- max(data[[paste0(y_var, "_avg")]]) * limit_factor
ggplot(data, aes(x = hour, y = get(paste0(y_var, "_avg")), fill = am_pm)) +
geom_bar(stat = "identity", position = "dodge") +
scale_fill_manual(values = color_palette) +
labs(title = paste("Average", y_label, "by Hour"),
x = "Hour",
y = paste("Average", y_label)) +
guides(fill = guide_legend(title = NULL)) + # Remove legend title
theme_minimal() +
theme(panel.grid.major.x = element_blank()) +
coord_cartesian(ylim = c(0, y_limit))
}
# Assuming your dataset is named 'summary_data'
data <- summary_data
# Create the bar plots with dynamically adjusted y-axis limits
calories_plot <- generate_bar_plot(data, "calories", "Calorie Burn", limit_factor = 1.2)
total_intensity_plot <- generate_bar_plot(data, "total_intensity", "Total Intensity", limit_factor = 1.2)
steps_plot <- generate_bar_plot(data, "step_total", "Steps Taken", limit_factor = 1.1)
# Print the plots
print(calories_plot)
print(total_intensity_plot)
print(steps_plot)
Observations:
Intensity: On average, users engage more actively at 5:00 AM, 8:00 AM, 5:00 PM, and 7:00 PM.
Step Count: On average, users record more steps at 8:00 AM and 7:00 PM.
These observations suggest that user activity may be influenced by their daily routines and responsibilities, with higher activity levels before or after typical workday hours.
# Function to create a box plot with customizable orientation and colors for a given y-axis column
create_custom_box_plot <- function(data, x_axis_col, y_axis_col, orientation = "ver", colors) {
ggplot(data, aes_string(x = x_axis_col, y = y_axis_col, fill = x_axis_col)) +
geom_boxplot(width = ifelse(orientation == "ver", 0.2, 0.5),
color = "black") + # Remove fill = "white"
labs(x = "", y = y_axis_col,
title = paste("Comparison", x_axis_col, "for", y_axis_col)) +
scale_fill_manual(values = colors) +
guides(fill = "none") + # Remove the legend for fill color
theme_minimal() +
if (orientation == "hor") coord_flip()
}
# Call the function with custom colors as a tuple and specify the x-axis label
custom_colors <- c("#FFA500", "#ADD8E6") # Orange for AM, Light Blue for PM
for (col in c("calories", "total_intensity", "step_total")) {
plot <- create_custom_box_plot(hourly_df, x_axis_col = "am_pm", y_axis_col = col, orientation = "hor", colors = custom_colors)
print(plot)
}
AM vs. PM Activity: The “am_pm” column indicates whether the activity occurred during the morning (AM) or afternoon/evening (PM). You can compare the activity patterns between these two periods and explore any differences in user behavior during these times.
EDA for seconds_heartrate_clean
str(seconds_heartrate_clean)tibble [2,483,658 × 3] (S3: tbl_df/tbl/data.frame)
$ id : chr [1:2483658] "2022484408" "2022484408" "2022484408" "2022484408" ...
$ date_time : POSIXct[1:2483658], format: "2016-04-12 07:21:00" "2016-04-12 07:21:05" "2016-04-12 07:21:10" ...
$ heart_rate: num [1:2483658] 97 102 105 103 101 95 91 93 94 93 ...seconds_heartrate_clean %>% summary() id date_time heart_rate
Length:2483658 Min. :2016-04-12 00:00:00.00 Min. : 36.00
Class :character 1st Qu.:2016-04-19 06:18:10.00 1st Qu.: 63.00
Mode :character Median :2016-04-26 20:28:50.00 Median : 73.00
Mean :2016-04-26 19:43:52.24 Mean : 77.33
3rd Qu.:2016-05-04 08:00:20.00 3rd Qu.: 88.00
Max. :2016-05-12 16:20:00.00 Max. :203.00 n_distinct(seconds_heartrate_clean$id)[1] 14# Group by 'id' and calculate the average heart rate for each user
average_heart_rate <- seconds_heartrate_clean %>%
group_by(id) %>%
summarise(average_heart_rate = mean(heart_rate, na.rm = TRUE))
# Bar Plot with smaller bars and custom breaks on the heart rate axis
bar_plot <- ggplot(average_heart_rate, aes(x = id, y = average_heart_rate)) +
geom_bar(stat = "identity", fill = "#ADD8E6", width = 0.8) + # Adjust the width here (e.g., 0.5 for smaller bars)
labs(x = "User ID", y = "Average Heart Rate", title = "Average Heart Rate for Each User") +
theme_minimal() + coord_flip() +
scale_y_continuous(breaks = seq(0, 100, 10), limits = c(0, 100))+ # Set custom breaks and limits for the y-axis
geom_hline(yintercept = 60, linetype = "dashed", color = "green") +
geom_hline(yintercept = 100, linetype = "dashed", color = "green")
# Display the bar plot
print(bar_plot)
Observations:
- Users’ average heart rate is within the normal range. Nothing remarkable here.
- We could suggest adding a feature to the app that sends notifications to users whose average resting heart rate falls outside the normal range.
EDA for weight_logs_clean
str(weight_logs_clean)tibble [67 × 8] (S3: tbl_df/tbl/data.frame)
$ id : chr [1:67] "1503960366" "1503960366" "1927972279" "2873212765" ...
$ date_time : POSIXct[1:67], format: "2016-05-02 23:59:59" "2016-05-03 23:59:59" "2016-04-13 01:08:52" ...
$ weight_kg : num [1:67] 52.6 52.6 133.5 56.7 57.3 ...
$ weight_pounds : num [1:67] 116 116 294 125 126 ...
$ fat : num [1:67] 22 0 0 0 0 25 0 0 0 0 ...
$ bmi : num [1:67] 22.6 22.6 47.5 21.5 21.7 ...
$ is_manual_report: logi [1:67] TRUE TRUE FALSE TRUE TRUE TRUE ...
$ log_id : num [1:67] 1.46e+12 1.46e+12 1.46e+12 1.46e+12 1.46e+12 ...Fat:Body fat percentage recorded.
BMI: Measure of body mass index based on the height and weight in the participant’s Fitbit.com profile.
isManualReport: If the data for this weigh-in was done manually (TRUE), or if data was measured and synced directly to Fitbit.com from a connected scale (FALSE).
For more detail check : Fitbit data dictionary
weight_logs_clean %>% summary() id date_time weight_kg weight_pounds fat
Length:67 Min. :2016-04-12 06:47:11.00 Min. : 52.60 Min. :116.0 Min. : 0.0000
Class :character 1st Qu.:2016-04-19 15:19:45.00 1st Qu.: 61.40 1st Qu.:135.4 1st Qu.: 0.0000
Mode :character Median :2016-04-27 23:59:59.00 Median : 62.50 Median :137.8 Median : 0.0000
Mean :2016-04-27 15:39:54.27 Mean : 72.04 Mean :158.8 Mean : 0.7015
3rd Qu.:2016-05-04 15:24:10.50 3rd Qu.: 85.05 3rd Qu.:187.5 3rd Qu.: 0.0000
Max. :2016-05-12 23:59:59.00 Max. :133.50 Max. :294.3 Max. :25.0000
bmi is_manual_report log_id
Min. :21.45 Mode :logical Min. :1.460e+12
1st Qu.:23.96 FALSE:26 1st Qu.:1.461e+12
Median :24.39 TRUE :41 Median :1.462e+12
Mean :25.19 Mean :1.462e+12
3rd Qu.:25.56 3rd Qu.:1.462e+12
Max. :47.54 Max. :1.463e+12 # Create boxplots for "bmi" and "weight_pounds"
columns_to_analyze <- c("bmi", "weight_pounds")
# Call the function to create boxplots
create_boxplots_in_one_output(weight_logs_clean, columns_to_analyze, decimal_places = 2)
entry_count <- weight_logs_clean %>%
group_by(id, is_manual_report) %>%
summarize(entry_count = n(), .groups = "keep") %>%
arrange (- entry_count)
print(entry_count)NA
NA# Check users that reported fat percentage
weight_logs_clean %>% filter(fat != 0)Observation: - Only two users reported fat percentage
# Calculate total entries
total_entries <- sum(entry_count$entry_count)
average_bmi_weight <- weight_logs_clean %>%
group_by(is_manual_report) %>%
summarize(mean_bmi = mean(bmi, na.rm = TRUE),
mean_weight_pounds = mean(weight_pounds, na.rm = TRUE),
entry_count = n(),
entry_count_percentage = round((n()/total_entries)*100,2),
.groups = "keep")
print(average_bmi_weight)NAObservation: - 61% of users log their weight manually, while 39% sync their weight from other devices
create_custom_box_plot(weight_logs_clean , x_axis_col = "is_manual_report", y_axis_col = "weight_pounds", orientation = "hor", colors = custom_colors)
# Use the remove_outliers function we created previously to remove the outliers
columns_with_outliers <- c("weight_pounds", "bmi")
# Loop through each column and remove outliers
for (col in columns_with_outliers) {
weight_logs_clean <- remove_outliers(weight_logs_clean, col)
}dim(weight_logs_clean)[1] 64 8create_custom_box_plot(weight_logs_clean , x_axis_col = "is_manual_report", y_axis_col = "weight_pounds", orientation = "hor", colors = custom_colors)
Observations:
It appears that users who log their weight data manually have a lower median weight than users who sync their weight from other devices.
The previous observation should be viewed as exploratory and could benefit from additional data. The weight log dataset only has 68 entries; more data would be needed to evaluate these hunches.
The lack of completeness in the weight log dataset could indicate a lack of user engagement.
next
- Revise notebook -Check for grammar
- Organize references and check headings and formatting
- Write final recommendations _ Write limitations
- Work on report and final recommendations
convert from rmarkdown to github
https://mbounthavong.com/blog/2022/7/30/hosting-a-r-markdown-html-file-on-a-github-page
Reference:
- EDA
- Histograms
- Plotting histograms with ggplot2,
- Error bars vs CI
- [Add density line to histogram](https://r-coder.com/density-plot-r
- Categorical, ordinal, interval, variables
Insights and recommendations
Another source:
https://www.kaggle.com/datasets/arashnic/fitbit/discussion/313589?search=data
#paper https://dl.acm.org/doi/pdf/10.1145/3339825.3394926
this is it: physical innactivity. Plot a barplot with percentages. https://www.cdc.gov/physicalactivity/data/inactivity-prevalence-maps/index.html#Race-Ethnicity
information about physical activity guidlines (sex and age):
https://www.cdc.gov/nchs/products/databriefs/db443.htm Elgaddal N, Kramarow EA, Reuben C. Physical activity among adults aged 18 and over: United States, 2020. NCHS Data Brief, no 443. Hyattsville, MD: National Center for Health Statistics. 2022. DOI: https://dx.doi.org/10.15620/cdc:120213
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